How AI Makes You a Better PM ?
Artificial intelligence makes you a better project manager. Just not in the way that you think.
Like many, I have been exploring the world of generative AI and what it is useful for. Like many, I have very mixed opinions. There are times where it appears brilliant and entirely appropriate. There are also times where it feels like using AI involves a lot of hard work for questionable value. In particular, there are days where those thoughts happen in the same session.
Many of us are exploring how to make the use of generative AI relevant to our jobs and our lives. In the breathtakingly short span of time since tools like ChatGPT have been available, there has been wide recognition that there is opportunity here. The ability to engage in something that feels like conversation and get meaningful content seems just this side of magic. As a participant mentioned in a presentation yesterday, it’s like having “a little buddy at your side to help you out.”
But help you with what? This has been the ongoing question. I know of instances where people have looked for help in writing reports. Others have asked for templates or outlines of how to write a deliverable or a presentation. There have been attempts to use generative AI to draft deliverables (sometimes in their entirety; in other instances a section at a time). AI has been used to summarize extensive content to its essential essence. It helps to generate Excel formulas and code snippets.
All this is helpful, as far as it goes. While there is no question that AI can feel like a supportive little buddy, it very often feels more like a lazy, recalcitrant and belligerent buddy. For all that you can prompt the creation of worthwhile content, there are also times where you get far less than you hoped for. Shortcuts get taken, responses are remarkable in their brevity and entire instructions get completely ignored.
The more I’ve used generative AI, the more convinced I have become that it can make you a better project manager. There are very real benefits to learning and using AI to advance your abilities. They just aren’t the ones you expect to find on the surface. One of the ways has involved exploring how to generate example templates and deliverables that can illustrate what good examples of projects look like.
The start of this investigation was an exercise in frustration. I crafted prompt after prompt, trying to get the AI to produce consequential results that were actually good examples of the deliverable in question. Very often, they were not. Sometimes, all I would get was a template, without any example content whatsoever. In other instances, what I would get would be exceedingly minimal. Essential points of content, but without any elaboration or context.
This persisted for longer than I would like to admit. Over time, though, I learned how to structure the prompts I was creating to better refine the results I was looking for. I started to get outcomes that were more in line with what I was looking for. Moving sentences forward or backward in the prompt changed their emphasis (or risked them being ignored entirely). Changing the tense of statements made a profound difference. Crafting short, simple, declarative sentences helped a great deal.
Repetition was a factor as well. Reiterating the same thing in subtly different ways helped to reinforce key points. So did being specific about the details of what was being sought. Asking for sample risks for an example project might result in two or three. Specifying the number of risks resulted in more. Asking for 20 sample risks meant I might get seven or 10 identified (but the AI would still get lazy and tap out, never generating more than 10).
The most interesting situation was when I asked the AI to generate a resource plan with appropriate estimates. I expected a summary of people that would work on the project, with corresponding estimates of effort and commitment. What I got instead was another budget, with expenditure categories and financial projections. If I squinted, I could see how my request got interpreted that way. It was certainly not what I intended, but you could argue that the end product fit the letter of what was being requested.
I eventually got the effort estimates that I was seeking, but it took some work to get there. A great deal of refinement of the prompt shifted the focus from stuff to people, and from dollars to hours.
领英推荐
It was here that I realized where artificial intelligence can genuinely help to improve project management. It has nothing to do with the results that the AI produces. It has everything to do with the specificity of the prompts that you use to generate those results.
AI is making me better at what I do because I am needing to get much clearer and more precise about the results that I am looking for. I am looking at the words I write through a very different lens, thinking about how they might be interpreted (and misinterpreted).
For all that I consider myself to be a good writer now, I am learning to write more precisely and explicitly. Despite decades of articulating requirements and setting expectations, I am developing even greater clarity.
In the coming years, generative AI is going to continue to be refined. It will inevitably produce more meaningful and relevant results. The answers that it provides and the output that it produces will be that much richer and more useful. Given how much the technology has evolved in only the two years since it has been generally available, it is possible to imagine exponentially improved results in the future.
What will make the difference between good and exceptional results, however, will be the precision by which questions are asked and prompts are crafted. The value of framing good requirements is something that is amply demonstrated by interacting with AI. The benefits of doing so, however, extend far beyond.
For me, the principal value of AI is its ability to hone and refine the critical thinking necessary to frame good requirements. It is a low-stakes, consequence-free means of testing how clear and specific you are about the results that you are looking for. You get feedback in real time, and the opportunity to continue refining and clarifying until you get the response that you intended.
Where that skill is most particularly relevant is not in interacting with AI. The real value of building clarity is in how you interact with your project team, your clients and your stakeholders. When we communicate with others, there is all too often room for misunderstanding or misinterpretation. The clarity we learn in communicating with generative AI is one that directly translates into setting better expectations and articulating more refined requirements when interacting with others.
Lessions Learnt: AI will make you a better project manager not by giving you the answers, but through helping you to ask better questions and express clearer expectations.